News & Views | Published:


Death and methylation

Naturevolume 409pages141143 (2001) | Download Citation


Malignant melanoma cells can resist committing suicide when attacked by chemotherapy. The explanation lies in the discovery that a key gene in the cell-death pathway is switched off in this cancer.

An ability to avoid committing suicide is one of the keys to a cancer cell's survival, whether it is spreading from one part of the body to another by metastasis or facing attack from chemotherapy. Malignant melanoma is a particularly nasty form of cancer in this respect. It is both highly metastatic and resistant to chemotherapy, implying that it has mastered the art of avoiding suicide by not implementing the biochemical pathway that leads to cell death (apoptosis). How melanomas do this was a mystery until now, because — unlike other cancer cells — melanoma cells usually have fully functional p53 genes, which are important in triggering apoptosis.

Writing on page 207of this issue1, Soengas and colleagues go a long way to solving the riddle. They show that malignant melanoma cells can avoid suicide by inactivating a gene at a step further on from p53 in the apoptosis pathway. This means that they can survive during chemotherapy and metastasis even though their p53 genes are functioning normally. But what makes this study doubly interesting is that the inactivated gene — Apaf-1 — is merely switched off, instead of being completely lost or mutated. The implication is that the suicide pathway could potentially be reactivated in melanomas, making them less dangerous.

The p53 protein is a key player in the cell-suicide pathway2,3 (Fig. 1). When a cell suffers DNA damage or certain types of stress, its p53 proteins are activated and send a death signal to the cell. p53 is kept in check by a protein called MDM2, which causes the destruction of p53. MDM2 itself is under the control of the p14ARF protein, which confines MDM2 to a subsection of the nucleus (the nucleolus). This prevents p53 breakdown. Stress signals are transmitted to p14ARF in part through a death-associated protein kinase, DAPK4. This apoptosis pathway is disrupted in most human cancers2,3 by downregulation or loss of p14ARF, upregulation of MDM2, or mutation (and therefore inactivation) of p53.

Figure 1: The p53-mediated death pathway, and how it can be subverted by DNA methylation.
Figure 1

Increased levels of p53 can lead to the initiation of a cascade of events that results in cell death by apoptosis. p53 is kept in check by a pathway involving DAPK, p14ARF and MDM2. This pathway is disabled in some way in most human cancers, often by inactivation of p53. Malignant melanomas, however, often have intact p53 genes. As shown by Soengas et al .1, these cancers do not express one of the downstream members of the cascade — Apaf-1 — and thus escape apoptosis in response to stresses such as chemotherapy. The Apaf-1 gene seems to be disabled by abnormal methylation. This has also been observed for the DAPK and p14ARF genes, and DNA methylation is involved in generating mutations in the coding region of the p53 gene in other cancers. So DNA methylation can contribute in several ways to inactivating the apoptosis pathway.

The exact mechanisms by which p53 promotes apoptosis are not known, but probably involve Bax, a molecule that stimulates the release of cytochrome c from mitochondria. Cytochrome c activates the Apaf-1 protein, which in turn activates the enzyme procaspase-9, resulting ultimately in cell death. Several proteins can interrupt this death cascade and thus keep the cell alive. The existence of such a complex pathway (with many feedback loops not shown in Fig. 1) certainly gives plenty of places at which death can be subverted, to which we can now add the inactivation of Apaf-1.

What makes the results of Soengas et al.1 so interesting is that they reveal a reversible switching off — rather than a permanent deactivation — of the Apaf-1 gene. The switch may involve the addition of methyl groups to cytosine nucleotides in DNA, and the removal of acetyl groups from the histone proteins that bundle up DNA into the compressed form seen in the nucleus. Soengas et al. show that the Apaf-1 gene can be turned back on by treating cultured melanoma cells with inhibitors of DNA methylation or histone deacetylation.

This 'epigenetic' gene silencing is increasingly being recognized as a common way in which cancer cells inactivate cancer-related genes5,6. Attention has focused on the methylation of cytosines in genes with a high proportion of cytosine–guanosine dinucleotides (the methyl-acceptor sequence) in their promoter regions. After these 'CpG islands' are methylated, and after changes associated with histone deacetylation have occurred, the relevant genes become silent. One example is the abnormal methylation of the promoter of the p14ARF gene. This results in the downregulation of the gene and the degradation of p53, disabling the suicide pathway7. The promoter of the DAPK gene is also subject to silencing by methylation8.

But unusually Soengas et al. find that, although the Apaf-1 promoter is a CpG island (and therefore a potential target for methylation), there were no changes in the methylation of this region before and after melanoma cells were treated with the methylation inhibitor. So the exact mechanism by which methylation switches off this gene remains in doubt. Perhaps the methylation inhibitor reactivates an unknown gene that controls Apaf-1 , or perhaps it demethylates another control region of the Apaf-1 gene, such as an enhancer or insulator.

In any case, it seems that the methylation of cytosine nucleotides in human cancer cells can help to inactivate the apoptosis pathway at several points — either upstream (at p14ARF or DAPK) or downstream (at Apaf-1) of p53. And, as well as contributing to gene silencing, methylation can also substantially increase the occurrence of harmful single-nucleotide mutations when it takes place in the coding regions of genes such as p53 (ref. 9). We clearly pay a price for having methylated cytosine residues in our genomes. Methylation is essential in controlling gene expression under normal circumstances. But it also has several ways in which it can disable the pathways that protect us from cancer.

On the other hand, the fact that the Apaf-1 gene can be silenced by epigenetic changes yet reactivated by drug treatment may have clinical benefits. It might be possible to make melanoma cells sensitive to chemotherapy. Moreover, other important pathways — such as those involving retinoic-acid receptor β2 (ref. 10), needed for cellular differentiation, or the DNA-repair protein MLH1 (ref. 11) — can be abnormally silenced by DNA methylation. In fact, the list of genes subject to abnormal epigenetic downregulation is growing rapidly. The search for ways to reactivate them will, no doubt, be a major area of research — perhaps involving high-throughput screens such as gene-expression chips — over the next few years. But it may come as a surprise to many scientists that little of our detailed molecular knowledge of how cells function is actually being used at present in treating human cancers.


  1. 1

    Soengas, M. S. et al. Nature 409, 207–211 (2001).

  2. 2

    Soengas, M. S. et al. Science 284, 156–159 (1999).

  3. 3

    Vousden, K. H. Cell 103, 691–694 ( 2000).

  4. 4

    Raveh, T. et al. Nature Cell Biol. 3, 1– 7 (2001).

  5. 5

    Jones, P. A. & Laird, P. W. Nature Genet. 21, 163–167 (1999).

  6. 6

    Baylin, S. B. et al. Adv. Cancer Res. 72, 141– 196 (1998).

  7. 7

    Robertson, K. D. & Jones, P. A. Mol. Cell. Biol. 18, 6457–6473 ( 1998).

  8. 8

    Kissil, J. L. et al. Oncogene 15, 403–407 (1997).

  9. 9

    Rideout, W. M. III, Coetzee, G. A., Olumi, A. F. & Jones, P. A. Science 249, 1288– 1290 (1990).

  10. 10

    Widschwendter, M. et al. J. Natl Cancer Inst. 92, 826– 832 (2000).

  11. 11

    Herman, J. G. et al. Proc. Natl Acad. Sci. USA 95, 6870– 6875 (1998).

Download references

Author information


  1. USC/Norris Comprehensive Cancer Center, Keck School of Medicine of the University of Southern California, Los Angeles, 90089-9181, California, USA

    • Peter A. Jones


  1. Search for Peter A. Jones in:

Corresponding author

Correspondence to Peter A. Jones.

About this article

Publication history

Issue Date


Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing